import os import json import re import subprocess from docx import Document FILE_PATH = ".data/AKCES/AKCES 1/kontroly-RTF/" OUTPUT_FILE = ".data/hf_dataset/cnc_skript2012/test.jsonl" def convert_to_docx(file_path): """Convert a given file to DOCX using LibreOffice.""" new_file_path = file_path + 'x' command = ['libreoffice', '--headless', '--convert-to', 'docx', '--outdir', os.path.dirname(file_path), file_path] subprocess.run(command, check=True) return new_file_path def read_docx_file(file_path): """Extract text from a DOCX file.""" doc = Document(file_path) text = "\n".join([para.text for para in doc.paragraphs]) return text def main(): # Step 1: Convert all RTF and DOC files to DOCX for root, dirs, files in os.walk(FILE_PATH): for file in files: file_path = os.path.join(root, file) if file.endswith('.rtf') or file.endswith('.doc'): # if the docx file already exists, skip the conversion docx_file_path = file_path[:-3] + 'docx' if os.path.exists(docx_file_path): continue try: convert_to_docx(file_path) except subprocess.CalledProcessError as e: raise ValueError(f"Error converting {file_path}: {e}") # Step 2: Process all DOCX files to extract text output_data = [] for root, dirs, files in os.walk(FILE_PATH): for file in files: file_path = os.path.join(root, file) if file.endswith('.docx'): try: text = read_docx_file(file_path) # postprocessing text = text.replace("\xa0", " ") text = text.replace("\n \n", "\n\n") text = re.sub(r' +\n', '\n', text) text = re.sub(r' +', ' ', text) text = text.strip() output_data.append({ "text": text, "original_file": file[:-5], # remove the extension }) except Exception as e: print(f"Error reading {file_path}: {e}") os.makedirs(os.path.dirname(OUTPUT_FILE), exist_ok=True) # Write the output data to a JSONL file with open(OUTPUT_FILE, 'w', encoding='utf-8') as jsonl_file: for entry in output_data: jsonl_file.write(json.dumps(entry, ensure_ascii=False) + '\n') if __name__ == "__main__": main()